Expiration example:

• backshift --save-directory /backshift-production/save-directory --expire --check-hashes

Notes

• You can drop the --check-hashes for speed, but then file chunks will not be compared against their cryptographic digests. Zero-length chunks will still be detected without --check-hashes.
• The first time you run an expire, it'll ask you how long you want to keep a given file in the backshift repository. You currently must enter the time to keep files, in units of seconds. The author likes to use 1 year, which is about 31556925 seconds.
• backshift-expire is deprecated. Please use backshift --expire instead.

If you ever need to keep a single host's (or filesystem's) backup data longer than the retention interval allows, you can do one of the following:

• do a restore of the data you need to keep as a tar archive, optionally extracting it to another hierarchy
• if you cannot lose the full history:
  • rsync the entire repo to another directory hierarchy
  • change the retention interval for the entire repo - temporarily or permanently. This is a number of seconds in <repopath>/overall/maximum-age
  • turn off your expiration cronjob (or turn it off in your similar scheduling tool)